Blind image blur assessment by using valid reblur range and histogram shape difference
作者：
Highlights：
• Reference-free, training-free and transformation-free automatic blur assessment.
• A reblur algorithm is proposed to create valid reblur range and reblur images.
• Blurriness is estimated from the shape difference of local histograms.
• Experiment results indicate high correlation with human perception of blurriness.
• The proposed method performs better than recent methods in the spatial domain.
摘要
•Reference-free, training-free and transformation-free automatic blur assessment.•A reblur algorithm is proposed to create valid reblur range and reblur images.•Blurriness is estimated from the shape difference of local histograms.•Experiment results indicate high correlation with human perception of blurriness.•The proposed method performs better than recent methods in the spatial domain.
论文关键词：Blur metric,Valid reblur,Histogram shape,Spatial domain
论文评审过程：Received 9 October 2013, Revised 26 January 2014, Accepted 13 March 2014, Available online 26 March 2014.
论文官网地址：https://doi.org/10.1016/j.image.2014.03.003
